Skip to content

[https://nvbugs/6423991][fix] Qwen3.5-VL MoE DFlash compatibility #16090

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
amukkara wants to merge 1 commit into from
Closed

[https://nvbugs/6423991][fix] Qwen3.5-VL MoE DFlash compatibility #16090

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
13 changes: 13 additions & 0 deletions tensorrt_llm/_torch/models/modeling_qwen3vl.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1306,6 +1306,19 @@ def post_config(self):
def vocab_size_padded(self) -> int:
return self.llm.vocab_size_padded

# Forward spec-dec attrs to the inner causal LM (``self.llm``): this wrapper
# is not a spec-dec model but is the outer model those paths look them up on.
@property
def draft_config(self):
return getattr(self.llm, "draft_config", None)

@property
def draft_model(self):
return getattr(self.llm, "draft_model", None)

def load_draft_weights(self, *args, **kwargs):
return self.llm.load_draft_weights(*args, **kwargs)

def infer_max_seq_len(self) -> int:
return self.llm.infer_max_seq_len()

Expand Down
Loading